节点文献

基于事件注入技术的网络可信性评测研究

Research of Network Dependability Evaluation Based on Event Injection

【作者】 庞永刚

【导师】 王慧强;

【作者基本信息】 哈尔滨工程大学 , 计算机应用技术, 2007, 博士

【摘要】 计算机网络技术的发展及其应用领域的不断扩大使得网络系统体系结构变得越来越复杂,网络使用者对网络性能的要求也越来越高,因此网络系统的可信程度显得尤为重要。传统的可靠性评测已经不能体现网络的可信程度,在可靠性基础上,可信性概念得以不断完善,增加了可用性、防危性、安全性和可维修性等属性,如何对网络可信性进行评测则成为可信性研究的一个重点。可信性评测的内容包括网络系统硬件、软件和人为因素等,具有多属性、多指标等特点,这使得可信性评测变得更加困难。本文在系统全面地分析和总结相关领域最新成果的基础上,对可信性评测问题进行了深入地研究,主要进行了以下工作:第一,介绍了可信性及相关概念研究发展,即传统可信性概念研究、源于美国军方的可信赖计算研究、微软的高可信计算和可生存性的提出。概括了当前与网络可信性评测有关的研究和取得的研究成果,包括网络系统硬件可靠性评测研究、软件可靠性评测问题、网络安全评测研究和防危性研究取得的进展和成果,指出了当前可信性评测研究存在的不足。第二,在分析已有研究的基础上,建立了计算机网络可信性评测的指标框架,包括软硬件的可靠性、可用性、防危性、安全性和可维性,指标框架涵盖了网络可信性评测的全部属性和重要评测指标。第三,为了满足网络可信性评测的需要,提出并完善了用于网络可信性评测的事件注入模型。分析了事件注入模型与传统故障注入模型的异同及事件注入模型的优点,指出了模型的关键技术问题,最后设计了进行网络可信性评测的事件注入系统。第四,应用事件注入技术对网络系统可信性部分属性进行了评测实验研究,实验结果验证了事件注入模型的有效性与可行性。第五,综合考虑网络可信性评测的多属性、多指标和多种因素的特点,引入了模糊综合评测方法进行网络可信性评测研究。目前网络可信性评测研究还处于不成熟阶段。本文旨在建立完整、可行的可信性评测指标框架,提出新的网络可信性评测模型,并对实验方法和方案进行研究。本文的工作可以为网络可信性评测研究奠定理论和实验基础。

【Abstract】 Development of network technology and expansion of network applications make the architectures of network system become more complicated and users’ requirement on network much higher, and hence the trustiness degree of network systems shows more importance. Traditional reliability evaluation can not reflect the trustiness degree of network any longer, how to evaluate network dependability has become an emphasis in dependability researches. Besides of reliability, the concept of dependability is further perfected. The attributes of availability, safety, security and maintainability are added into it, which make the dependability evaluation more difficult. The objects of dependability evaluation include network hardware, software and human factors, and the evaluation is of multi-attribute and multi-index features. After systematically studying and summarizing the newest research, this paper deeply studies the dependability evaluation, the work are mainly listed as follows:Firstly, development of research on dependability and its related concepts is introduced, including concepts of dependability, trusted computing, trustworthy computing and survivability. Research and findings related to dependability evaluation are summarized, including research progress or findings on hardware reliability evaluation in network system, software reliability evaluation, network security and safety evaluation, and then shortcomings of current research on dependability evaluation are pointed out.Secondly, by analyzing the existing researches, dependability evaluation index framework of computer network is constructed, which includs reliability, availability, safety, security and maintainability of hardware and software, and this framework contains all attributes and important evaluation indexes of network dependability evaluation. Thirdly, to meet the requirements of network dependability evaluation, event injection model used in network dependability evaluation is proposed and consummated. Differences between event injection model and traditional fault injection model and the advantages of event injection model are analyzed, key technologies are proposed and event injection system used to evaluate the network dependability is designed.Fourthly, applying event injection technology, experimental researches on evaluating of some attributes of network system dependability are carried out. Experimental results validate that the event injection model is valid and feasible.At last, considering the influences of multi-attribute, multi-index and multi-factors on network dependability evaluation, fuzzy comprehensive evaluating is introduced into the network dependability evaluation.In a word, research on network dependability evaluation is in an immature stage, and this paper intends to construct an integrated and feasible dependability evaluation index framework, proposes a novel network dependability evaluation model, and carries out some research on experimental schemes. The work done in this paper can set up the foundation for future network dependability evaluation.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络